https://repository.hneu.edu.ua/handle/123456789/140 2026-07-17T11:08:16Z https://repository.hneu.edu.ua/handle/123456789/41430 Назва: Evaluation frameworks for large language models Автори: Знахур С. Короткий огляд (реферат): The paper addresses the challenge of assessing quality in large language model (LLM)-based systems, whose stochastic, non-deterministic outputs make classical pass/fail testing insufficient. The author proposes and empirically validates a structured, multi-method quality-assurance (QA) framework that combines four complementary evaluation strategies — an LLM-as-a-Judge holistic rubric, GEval reference-based chain-of-thought scoring, AspectCritic binary aspect-level verification, and SelfCheckBERTScore semantic-consistency measurement across repeated samples. The framework is applied to a locally hosted LLL relation model run through the Ollama runtime and evaluated on a purpose-built dataset of 500 prompts spanning factual, relational, and open-ended analytical categories. Pass rates across the four metrics range from 62% (GEval) to 76% (SelfCheckBERTScore), with performance declining consistently from factual to relational to analytical prompts — a gradient the author attributes to measurable prompt-level features such as length, number of referenced entities, and the presence of contrastive or negation cues. Pairwise inter-metric agreement, computed with Cohen's Kappa, shows substantial agreement between content-oriented metrics (LLM-as-a-Judge vs. AspectCritic, κ = 0.61; AspectCritic vs. GEval, κ = 0.65) but only fair agreement between SelfCheckBERTScore and the correctness-oriented metrics (κ = 0.35–0.43), confirming that output consistency and content correctness are structurally distinct quality dimensions. Manual failure analysis identifies three primary failure patterns — partial incompleteness (46% of failures), factual imprecision (32%), and semantic inconsistency (21%) — each detected preferentially by a different subset of metrics. Based on these findings, the author proposes three targeted, low-cost improvements: structured prompt templates that force explicit enumeration of relational components, GEval-based post-generation filtering as an internal quality gate, and reduced sampling temperature for analytical prompts to curb cross-sample divergence. The study demonstrates that no single evaluation method captures the full spectrum of LLM quality concerns and contributes a reproducible, API-independent evaluation pipeline applicable to other LLM-based systems. 2026-01-01T00:00:00Z https://repository.hneu.edu.ua/handle/123456789/41429 Назва: AI-driven information systems development: LLM-based code generation approaches Автори: Знахур Л. Короткий огляд (реферат): The paper presents a systematic review and empirical investigation of large language model (LLM)-based code generation systems covering the period 2022–2026, together with a proposed hybrid architecture that combines Retrieval-Augmented Generation (RAG) with a Mixture-of-Experts (MoE) routing strategy. The author surveys the three principal model families used for code generation — autoregressive decoders, masked language models, and encoder-decoder architectures — and synthesises benchmark results for leading systems such as Claude 3.7 Sonnet, GPT-4o, Gemini 2.5 Pro, DeepSeek V3 and Llama 4 Maverick across HumanEval, SWE-Bench, MBPP and LiveCodeBench. A Python-based prototype of the proposed hybrid model is implemented using QZhou-Embedding for query representation, a FAISS vector index for retrieval over a 10,000-snippet Python code corpus, three specialised MoE expert modules (Python, JavaScript, and security analysis), and an iterative self-reflection loop that verifies generated code through AST parsing and chain-of-thought semantic checks. Evaluated on the full 164-problem HumanEval benchmark, the hybrid model achieves a Pass@1 of 85%, exceeding CodeLlama by 15 percentage points and GPT-4o by 11 percentage points, while also producing code with a higher pylint score (8.5/10), lower cyclomatic complexity (2.3), and a higher Maintainability Index (85) than both baselines. The retrieval subsystem is shown to add less than 1% to end-to-end query latency while the generation stage accounts for roughly 94%, and the system scales acceptably under concurrent load. The author identifies three recurring limitation patterns — retrieval noise, self-reflection runtime overhead, and limited support for low-resource programming languages — and proposes fine-grained candidate re-ranking, adaptive reflection termination, and multilingual corpus extension as mitigation directions. The findings support the conclusion that combining retrieval grounding, task-specialised routing, and iterative verification yields measurable improvements in functional correctness and code quality over both commercial and open-source baselines. 2026-01-01T00:00:00Z https://repository.hneu.edu.ua/handle/123456789/41272 Назва: Circle-intersection-based equal-chord partitioning algorithm for a planar parametric curve Автори: Frolov О. V. Короткий огляд (реферат): This paper addresses the problem of equal-chord partitioning of planar parametric curves, which consists of determining a sequence of points on a curve such that the Euclidean distances (chords) between consecutive points are equal. Equal-chord discretization is important in computer-aided design, CNC machining, robotics, and computer graphics, as it provides stable geometric and technological properties compared to parameter-uniform or arc-length-based sampling. This work aims to develop efficient algorithms for equal-chord partitioning of planar parametric curves in the classical setting, i.e., with fixed endpoints and a prescribed number of segments, while explicitly accounting for the multivalued nature of solutions arising from curve–circle intersections. The proposed approach is based on successive intersections of the curve with a moving circle of fixed radius. Starting from one or both endpoints, the circle center is placed at the current partition point, and subsequent points are obtained as intersections between the curve and the circle in a specified direction. To handle non-convex curves and curves with inflection points, all possible intersection solutions are organized into a tree structure, where each branch represents an alternative partitioning trajectory. The optimal partition is selected by minimizing the difference between the residual segment length and the circle radius. An experimental evaluation of planar Bézier curves of various orders demonstrates stable convergence of the proposed algorithms across a wide range of segment counts. As the number of segments increases, the multiplicity of curve–circle intersections decreases, and the computational complexity approaches linear behavior in practical scenarios. The proposed method avoids numerical arc-length integration and auxiliary grid-based approximations and can be naturally extended to spatial curves. 2026-01-01T00:00:00Z https://repository.hneu.edu.ua/handle/123456789/41105 Назва: Mathematical models for developing interactive components of distance education systems Автори: Karpenko M. Y. Короткий огляд (реферат): A formalized mathematical model of a dynamic economic system with regulated parameters is proposed to analyze economic equilibrium under the interaction of a producer, an aggregate consumer, and a price-regulating authority. The model accounts for production constraints, consumer utility, pricing dynamics, and the closed nature of the economic environment. It investigates supply–demand imbalances, identifies the conditions for restoring equilibrium through price adjustments, and examines the effects of unrealized stock and deferred demand. The dynamic framework enables the analysis of successive system states, evaluation of pricing strategies, and identification of critical operating regimes. Designed as the basis for an interactive simulator integrated into a distance learning environment, the model supports flexible parameter adjustment, enhances the realism of economic process simulation, and fosters analytical thinking and practical skills in economic modeling. 2026-01-01T00:00:00Z